Application of a hepatitis E transmission model to assess intervention strategies in a displaced persons camp in Uganda

Abstract

Hepatitis E virus is an enterically transmitted disease that mainly effects people in developing countries. It is known to spread through exposure to contaminated water sources (Balayan, 1997) and is also thought to spread through person-to-person transmission (Teshale et al., 2010). In recent years there have been outbreaks in internally displaced persons (IDP) camps in Sudan (Boccia et al., 2006; Guthmann et al., 2006) and Uganda (Siddiqui, 2010). In IDP camps interventions such as installation of hand washing points, water purification and latrine building are undertaken to control the spread of the disease. Hand washing points reduce the person-to-person transmission, water purification reduces exposure to the contaminated water source and latrine building reduces shedding of the virus by infected individuals to the environmental reservoir. Of interest is how effective these interventions are, which intervention should take precedence and how rapidly the interventions should be implemented. A differential equation based compartment model is developed for the transmission of hepatitis E via the faecal-oral route including both an environmental reservoir and person-to-person spread. Due to the long incubation and infective periods of hepatitis E, the model includes both exposed and infectious classes of individuals. Multiple exposed and infective classes are used to give more realistic gamma distributed infective periods. Where possible, parameters in the model are determined from knowledge of the aetiology of hepatitis E. The remaining unknown parameters are estimated by fitting to incidence data from four IDP camps during the recent Ugandan outbreak. This is performed using a Bayesian Markov chain Monte Carlo method where prior distributions for the unknown parameters are progressively refined until a posterior estimate of the parameter distributions is obtained. Consistent estimates for the parameter governing transmission from the environment to an individual are found for all four camps. Shedding to the environmental reservoir was more variable across the four camps. The person-to-person parameter showed the greatest variation across the camps. This is consistent with conflicting reports in the literature as to the importance of person-to-person transmission of hepatitis E and is clearly an area requiring further research. Before the model can be run to determine the effectiveness of interventions the timing of these interventions must be determined. A threshold number of cases need to be reported before interventions are considered and then there are various lags before interventions can be fully implemented. These lags include delays in testing, hardware delivery and installation and manpower and resource limitations. These delays are specific to each site and need to be determined in consultation with the field epidemiologists.